Value Add Baseball registered a quick strikeout of the New Republic’s Marc Tracy, who writes that Value Add blames the pitcher for unearned runs even though the printed formula does exactly the opposite (Strike 1). Tracy also asserts Value Add cannot be accurate because Clayton Kershaw is not ranked first, apparently not realizing the system he claims is superior ranks Kershaw even lower than Value Add (Strike 2). Finally, he goes back 54 years to pick a game that he believes best proves his case only to have it do the opposite (Strike 3).
Strike 1 – False Claim that Value Add blames the pitcher for unearned runs
STRIKE 1 – Tracy’s attempt to discredit Value Add rests on him in essence changing the formula, and then claiming it is wrong because of the change he made. The most amazing paragraph in Tracy’s piece reads:
Oh, and also, Value Add Baseball does not adjust for defenses! (Though it does adjust for ballparks.) “All runs are the responsibility of the pitcher,” (Dave) Cameron noted of the metric (indeed, it uses unearned runs, when in fact, such runs are considered “unearned” for a reason). “If you don’t adjust for defense, you’re missing part of the picture.”
This is simply a … misstatement.
Tracy can only make this assertion by completely reversing the formula, which is clearly laid out for all to see in the July 31 post:
To win, (the pitcher) needs to stay below an “ERA Needed” in each game, which is the total of: His team’s offensive runs that day MINUS unearned runs allowed by his defense MINUS relief runs allowed, DIVIDED BY his innings pitched TIMES nine. (emphasis added)
The unearned runs are absolutely blamed on the defense, and therefore show that the pitcher had a lower chance of winning the game. In essence, Tracy erases the words, “MINUS unearned runs allowed by the defense” and then claims that the formula is wrong because the words “MINUS unearned runs allowed by the defense” are not included, and that’s just dishonest.
Strike 2 – Value Add cannot be accurate because Kershaw is not ranked No. 1
Tracy asserts Value Add cannot be as accurate as WAR on Fangraphs because it does not have Kershaw as the best pitcher – a very funny argument since Value Add ranks him as the second most valuable while two WAR formulas on Fangraphs calculate that Kershaw is either the third or fourth most valuable pitcher. Tracy’s statement reads:
“Conservatives want their numbers. Even if it means that the Dodgers’ Clayton Kershaw–who is having an all-time season–is ranked the second-best pitcher this year…”
I have stated in posts that Kershaw is the best pitcher in baseball, so my not “wanting” him to be first is inaccurate and irrelevant. I do not doubt he is the best pitcher, but WAR and Value Add claim to measure VALUE, which is not always the same. This season, Pat Corbin has been slightly more valuable, as laid out in a side-by-side (one minor correction to the ERA Needed calculations in that post is that 0.1 innings pitched and 0.2 innings pitched have since been corrected to 0.33 and 0.67 in the program, but this does not impact the order of the rankings and could make very slight adjustments to value calculations.) Two of the Fangraph’s measurements indicate that Felix Hernandez and Adam Wainwright have a higher WAR than Kershaw, and one of the two methods also show Matt Harvey with a higher WAR to push Kershaw to No. 4 in value.
The very evidence that Tracy presents shows Value Add to be more accurate–if we accept his premise that Kershaw should be first.
Strike 3 – Haddix Appearance on May 29, 1959
Tracy goes back 59 years to find a game to discredit Value Add – Harvey Haddix’s 12 innings of perfect baseball on May 29, 1959 in which his team never scored and he finally lost. Tracy’s claim is that Value Add is unfair because it does not credit him with a victory in that game. Granted, just as traditional stats are unfair because Haddix did not get a “Win” for that outing, but that does not mean we should pretend he won the game and go back and adjust his record for the season.
Ironically Tracy has chosen a perfect example of a game in which Value Add succeeds and WAR, which Tracy is advocating, does not. Both Value Add and WAR make the same claim – they are the best measurement of how many MORE WINS a team has because of the player.
How many more wins does Value Add claim the Pirates won on May 29, 1959 because Haddix was pitching? Zero. How many extra wins does WAR calculate that the Pirates won on May 29, 1959 because Haddix pitched? About one. Which is correct? ZERO – the fact is that the Pirates did not win on May 29, 1959, so if a team has zero wins they cannot have one more win than they would have had with a replacement pitcher on the mound. In the very game Tracy picks, Value Add was correct and WAR was wrong in what both systems claim to do.
Certainly other systems such as a pitcher’s “Game Score” can measure how great an outing that was, but WAR and Value Add claim to measure how many more wins the pitcher gets his team, and it is easy to prove that in Tracy’s very example Value Add was correct. Value Add calculates that any pitcher given a zero percent chance of winning a game (needed below a 0.00 ERA for the game to win) is given a 0.2 bonus to the Raw Value Add score and that over the course of the season that puts all pitchers on equal footing despite their level of support.
Luckily, over the course of the season every pitcher gets opportunities to win in most outings. A pitcher who is given enough support to have a 30% chance to win 10 times and actually gets the team a 5-5 mark has added two wins for his team. He is just as VALUABLE as a pitcher who gets enough support to have a 50% chance to win 10 times and gets the team a 7-3 mark.
Tracy and Cameron (whose work really is fantastic) both acknowledge that WAR has weaknesses as well, but these weaknesses are much more prevalent pertaining to starting pitchers than on other positions. Measuring support by saying a team gives a pitcher 5.0 runs of support can be very deceiving. A good pitcher who gets exactly five runs of support every game will usually go 3-1 every four games. A pitcher who gets 20 runs of support one game and zero runs the next three games will almost certainly go 1-3. The first pitcher has received far greater support, as measured by Value Add. The “average runs of support” used in other systems calculates both pitchers received just as much support, which is inaccurate.
As always, make sure to follow our team on Twitter for breaking news and thoughts. I can be followed @jpudner, Tony Lee @TheTonyLee, Cole Muzio @ColeMuzio, Cameron Scales @CamScales, and Benjamin Chance @CBenjaminChance. Also, remember to follow our site @BreitbartSports.